Some existing data services rely on data collected by a plurality of computing devices to adjust output. The data services process such “crowdsourced” data to produce output for distribution to and consumption by various devices. For example, some existing location services rely on crowdsourced beacon fingerprint data to determine a location of a requesting device. The beacon fingerprint data includes a set of beacons observed at a particular location.
To help ensure the integrity of the crowdsourced data, some existing data services provide authentication to confirm the identity of the computing device submitting the data and encryption to prevent tampering with the submitted data. Authentication and encryption, however, cannot validate or vet the data contained within the submission to prevent the introduction of erroneous or malicious data into the data maintained by the data services. For example, a malicious device may associate the same set of observed beacons with multiple different locations. Crowdsourced data from this malicious device will pass authentication and encryption tests at the location service, yet will contaminate the data maintained by the location service reducing the location determination accuracy of the location service.